Dynamic Parameters Identification for 6-DoF Industrial Robots

Abstract

The 6-degree-of-freedom (DoF) industrial robot has been widely used in industrial production. Accurate dynamics model is essential for achieving accurate torque control, and dynamic parameters identification is fundamental for establishing a precise dynamics model. Furthermore, the influence of joint friction on joint torque is not negligible. Thus, in this paper, a linearized dynamics model including joint friction is established. Then, in order to sufficiently stimulate the dynamic characteristics, the excitation trajectory of each joint is designed by optimization criterion. The actual torque and motion data of each joint are collected from controller, and the parameters are identified by least square (LS) estimation method. The identified dynamic parameters in this paper are applied to construct a linearized dynamics model, and root mean square error (RMSE) values of two verification trajectories between the estimated joint torques of the dynamics model and the measured joint torques are less than the allowed values.